This is another reply from Richard that can help you:
>I really like conspiracy theories - did you have to dispell
>that one so soon?
>
>I’m really confused about the GF3-volume texture issue as
>I’ve been told flat out by many people from nVidia (possibly
>even yourself at the Gathering) that they were not
>supported, and to wait for a future card.
Agreed. At the time of The Gathering we didn’t expect to be able to enable
volume textures for GF3 in any reasonable timeframe. You can see from how
long it’s taken us to get our current DX8 driver out there that this isn’t a
totally unreasonable position.
Sometimes the things you’ll hear from us will seem to contradict something
which we’ve said at an earlier date. To a certain extent “that’s life”.
All we can do is give you the best information which is available at the
time - and things like this can change. The volume texture issue is one of
those. NVIDIA’s decisions can change in the light of new information that
we get about our chip yields, driver team load, market pressure, developer
demand etc.
>At the end of the day though I’ll just be happy for it to work
>- but I am worried now that there may be issues when they are
>used - like a texture stage disappears or what-not???
Don’t worry. A volume texture consumes a single texture stage. End of
story.
>As for TruForm you have to admit it does look good from a
>first glance and could you say why exactly you see this as
>a ‘dead-end’? Do you see subdivision type surfaces in
>graphics cards as a no-go generally?
Well, we actually see subdivision surfaces as the main thing to aim for. I
can’t go into details about how long we expect it to take, but it’s a
serious goal of ours for consumer graphics. You can tell that from the
amount of research papers which we publish from the likes of Henry Moreton
and Dave Kirk.
This is a much more useful line of development.
And for subdivision surfaces to be really useful they should really support
all of the neat features [creases, darts, arbitrary valences etc.] Getting
that into hardware will take some serious work from some of the best
graphics theoreticians out there.
>Personally, so far I’m glad to see that patches have been the
>first to make it to hardware as it makes most sense due to
>their support in modelling programs.
Agreed. This is an important first step.
>Oh, and while I may have your attention - in future can we
>please not loose a texture stage if we turn on a clip plane -
>please!
Well, yes… Right now the reason we’ve removed the clip plane support is
precisely because of this. The API is at it’s most useful when it’s
orthogonal.
[Strictly speaking the clip plane support is still present in this driver,
we just don’t expose the relevant caps. And sooner or later you can expect
this kind of support to be removed for exactly the reason you state.]
Richard “7 of 5” Huddy
Developer Relations, NVIDIA Corporation.